NASAg03 said:

---

bmks270 said:

---

NASAg03 said:

---

Okay here's my POV after working 5 years designing deep sea robotic vehicles for Nauticus Robotics. My experience includes a trip to WHOI to study their latest ROVs and apply to our vehicles. Our last vehicle was rated to 3000m. There were not humans involved, precisely because designing ANY deep sea vehicle is hard. It's a more challenging environment than space, and as such, we're trying to replace humans deep sea because of the cost and risk.

All of our vehicles were flooded, and only specialized electronics were placed in very expensive titanium pressure housings using American material stock, fabricated in America, with high quality metals. All were machined from single stock of material, pipe or billet. Why? Because any void can cause an immediate failure.

Any electronics with pressure-rated components were flooded with oil, and pressure compensated. Slightly positive pressure ensures oil is pushed out of the system so that no seawater gets into the electronics. We place water sensors in all electronics so that if they get wet, the computer will immediately know and bring the sub to surface before housing is completely flooded.

Most housings are pressure-assisted seals, and only need enough fastener preload to set the seal. After that, water pressure assists the seal until the metal flanges compress and transfer the load.

The pressure-assisted front "hatch" on Titan isn't the big issue. That's typical and a low risk item. The acrylic window is the higher risk item, as plastics creep, age, and change more with time, pressure and UV exposure than most metals. The biggest issue is the composites...which also have plastic involved.

We use composites for our unmanned subsea vehicle, but only as unpressurized vehicle structure. I've never seen any one use composites for a subsea pressure housing. That applies to ROVs and AUVs. Making a human-rated composite pressure vessel is unheard of.

Even using them for internal pressure, such as COPV, they typically have a metal liner to prevent leaks and are still leak-before-burst designs. You can't get that with external pressure, as any leak in the structural walls typically is catastrophic.

The mating seal between the Ti end caps and the composite cylinder is an adhesive bond, which isn't a big deal as it looks like the end cap has a flange that transmits axially load into the cylinder and the adhesive isn't that structural. However, the issue with this design is, any uneven loading on that flange can cause localized failures, ripples, and delamination of the composite, most likely internally due to the pressure distribution. This results in high stress regions that will avalanche since it's in compression.

My guess is that's what happened. After a few cycles, the internal flange lip of the composite cylinder wall delaminated due to higher-stress from the dome. Depending on post-dive inspections (if they did any), the delamination might not have even been detected at STP due to relaxing of the joint.

I'm curious what kind of NDE they did after fabrication and every dive, and what areas of the sub they inspected. I'm also curious if they did any leak monitoring or stress measurements during missions at high criticality regions.

---

It's 5 inches thick carbon fiber.
It's not easy to inspect.
The engineer who wanted do scans of the structure was fired.

---

NDE of CFRP or the bond at the CFRP / Ti dome interface isn't hard. You can do ultrasonic or x-ray. Gaps due to delamination would easily show up. Hell, they could automate it in 2 degrees of freedom making passes axially and around the circumference with a simple jig.

If this wasn't happening, it's because the CEO didn't think it was worth the cost, not because it's hard. He was wrong and cutting corners in every way possible.

---

I don't consider those inspections easy.
X-Ray a submersible the size of a car and correctly interpret the result?
Ultrasonic inspection of every square inch of of 5 inch thick carbon fiber that is far outside of the depth capability of most inspection tools? Hard includes time consuming and tedious. Not necessarily impossible.

This paper reports being able to inspect a 2.36 inch thick panel. The sub was 5 inches thick. Are there established tools you know of that can get 5 inches, or even 3 inches of penetration?

Not that it was impossible, like I said, the CEO fired the guy who wanted to do the more thorough inspections and scans. The CEO wanted easier.

https://www.mdpi.com/2076-3417/11/4/1508

"""
The pulse-compression technique was successfully implemented in combination with a linear phased array system to inspect a 60 mm thick CFRP sample with artificial defects.
…
These preliminary results encourage further improvement to the system and to extend the analysis of the benefits of PuC-PA inspection to more challenging samples with higher attenuation or larger dimensions, where the pulse-echo method cannot assure enough sensitivity to detect deep defects.
"""

one MEEN Ag said:

---

NASAg,

Great points. One big giveaway is if they put a single layer of E-glass between the titanium and carbon fiber. Carbon fiber has galvanic corrosion issues with a huge assortment of metals. My understanding of non marine CFRP (carbon fiber reinforced polymer for those playing at home) use is to put a layer of E-glass composite down first to separate the metal from the carbon fiber. E-glass, while way weaker in tension isn't nearly as big of a risk of galvanic corrosion. Thats pretty standard for steel pipe repair if you're using carbon fiber.

If they didn't even do that, then they didn't even talk to a single engineer during the design and procurement phase.

---

Very true. They probably did that or glass beads with 0.01" diameter or whatever the adhesive needs to control bond line thickness. I'm guessing they did post-machining of the bond interfaces for better tolerances, and that would negate the use of a E-glass layer.

The primary sealing surface is probably the flange faces on the end caps since those are seeing the highest compressive stresses. That's where I would put the bond material and whatever bondline controller you need to separate the dissimilar materials.

We routinely use e-glass barrier plies between Gr/Ep skins and aluminum substructure. We lay down several plies (tow placement) and then finish machine to match the profile of the substructure to eliminate shimming.

Hell, we learned the importance of this way back on the F-117 30 years ago.

If they did not do this, they did not employ anyone who had any experience whatsoever in using Gr/EP composites.

yawny06 said:

---

Aerospace may be a little tighter on safety factors, not sure but I suspect they would have to be since weight is so critical for flight. That was his background and I am curious if that didn't play into his mindset at all.

---

A lot sportier. Factor of safety 1.5 for manned vehicles; 1.25 for unmanned.

eric76 said:

---

Hydraulic Press testing Carbon Fiber and other materials



I wonder how many people would watch that and still use carbon fiber for the hull.

---

Good comparison for the isotropic materials (polymers and metals), but the CFRP tube layup probably isn't representative of what they used for the Titan cylindrical hull.

A 2X2 twill weave is horrible in compression. They used at least hoop bands, and probably unidirectional axial fibers as well for compressive axial loads.

A better comparison would be unidirectional layers in the axial and hoop direction alternating. Based on the stresses (2X more in the hoop), you might want 2X layers in the hoop direction. Alternate to prevent buckling.

NASAg03 said:

---

bmks270 said:

---

It's 5 inches thick carbon fiber.
It's not easy to inspect.
The engineer who wanted do scans of the structure was fired.

---

NDE of CFRP or the bond at the CFRP / Ti dome interface isn't hard. You can do ultrasonic or x-ray. Gaps due to delamination would easily show up. Hell, they could automate it in 2 degrees of freedom making passes axially and around the circumference with a simple jig.

If this wasn't happening, it's because the CEO didn't think it was worth the cost, not because it's hard. He was wrong and cutting corners in every way possible.

---

You can NDI CFRP that thick by ultrasonic? Asking because I genuinely don't know, I normally deal with structures well under an inch because that's what we make. Items go through ultrasonic water jet inspection.

bmks270 said:

---

Stat Monitor Repairman said:

---

Anyone see this alleged text message transcript between the support vessel and the sub?

They might have had their first indication of a problem 19 minutes before the implosion.

Said they got an alarm on the system that acoustically monitors the hull for cracks. Then decided to make an emergency ascent. Sounds like they somehow did some emergency procedure that released the skids. The vessel didn't respond like they expected.

Then they released more ballast which involved putting all 5 people to one side until ballast fell of hooks.

Sound like they were hearing the hull cracking and trying to figure out what to do.

---

I think transcripts floating around are fake. I've seen at least two or three versions.

---

That was my initial reaction that the transcript was fake.

Just the way they were communicating sounded bush league.

Doesn't really sound like there was any sort of report and respond protocol. Or standard terminology like you would expect. It read like they were doing whatever came to mind like a couple of teenagers casually chatting over text.

But the content sounds plausible. But is surprising nobody has come out and confirmed it as true.

There could have been a large shear load on the adhesive due to material thermal property differences between the carbon fiber and Ti.

The temperature gradient was over 40F from exterior surface to interior (assuming min. 68F inside).

with everything else we've seen from the company I wouldn't be surprised if there were no official communication protocols, or if there were they were routinely disregarded. that said I would be surprised if they are the official transcript.

I don't think those transcripts are legit.

Everything I read early on said that comms were basically one way from the mother ship to the sub. The sub had limited capability to send messages back, certainly not to the extend of the supposed transcripts.

Perhaps not but this engineer points out several clues that might mean they are authentic. Although he stopped short of saying that. Nonetheless, if authentic, it's paints a grim picture that the sun was descending too fast. Once the alarms went off, they spent 19 minutes in horror trying to ascend in vain.





Edit to fix thumbnail/link

Was there more than one water tight compartment in this sub that could have conceivably flooded?

Stat Monitor Repairman said:

---

Was there more than one water tight compartment this sub that could have conceivably flooded?

---

Most likely the pressure-sensitive electronics were in sealed enclosures. Some batteries are pressure rated, some have to be sealed as well. If any of those flooded, the buoyancy and sub performance would be negatively impacted, but the human-rated volume should be fine.

In such a scenario, you detect a flooded housing and drop ballast to slowly raise to the top. It's possible they were at a poor orientation to drop ballast and did have to shift inside to get that to work.

But that still wouldn't necessarily cause the sub to implode.

But it's possible you had multiple failures at once considering how many corners they cut.

one MEEN Ag said:

---

"Using the stated thickness of 5inches, that's an axial compressive stress of 20ksi (and end cap stress flange stress), and a "hoop" compressive stress of 37ksi. Combine the two and that's 42ksi von Mises stress, which doesn't even apply to brittle materials."

You know the craziest part in my mind is that is a lot of pressure, but not an insurmountable pressure. Even if you wanted a 3x factor of safety that is still only 126ksi yield steel. You can get super austenitic stainless steels with high nickel content into the 160-170s ksi's now. Its not even unobtanium to get high strength stainless steels or nickel based alloys.

They've already written the book on submersible design, fabrication and inspection. To me this just continues to drive home how stupid this was to try to innovate on hull design.

---

Totally agree. No reason to use a CFRP overwrap other than cost savings, and maybe schedule and mass. I'm guessing it was driven by cost. Making high-strength steel work would involve syntactic foam or buoyant metal spheres, which is really expensive. Safer and easier to inspect, but much more expensive.

"You're dead anyway."

From https://www.insider.com/oceangate-stockton-rush-titan-submersible-bolted-in-dead-anyway-2023-7

Quote:

---

Moments after Weed, the "Expedition Unknown" host Josh Gates, and Rush were locked in the sub with no way out except from the outside, Weed asked Rush what would happen if the vessel had to suddenly make an ascent in an emergency situation and was nowhere near its mothership.

Weed told Insider that Rush said, "'Well, there's four or five days of oxygen on board, and I said, 'What if they don't find you?' And he said, 'Well, you're dead anyway.'"

---

Hypothetically speaking, if they detached the skids in some sort of an emergency procedure ...

What happens when they get to the surface and have to mate back up with the barge?

If there's no landing skids because they were dumped as ballast ... how was that supposed to work?

You'd have a cylindrical vessel rolling around on deck while somebody out there trying to unbolt the cap?

They thought of that problem…

Had a couple of these on board….

https://www.harborfreight.com/solid-rubber-wheel-chock-96479.html

That's what chocks are for.

I guess you'd send divers down and secure it to the deck of the barge before it was re-floated. That's probably what they do.

Of the pieces recovered from the sea floor, were any of these of the carbon fiber hull?

In the video, there were the titanium rings, but it didn't look like there was any hull attached to the rings.

https://www.businessinsider.com/titanic-sub-photos-suggest-carbon-fiber-hull-to-blame-expert-says-2023-6?amp

BQ78 said:

---

https://www.businessinsider.com/titanic-sub-photos-suggest-carbon-fiber-hull-to-blame-expert-says-2023-6?amp

---

So it appears that they have recovered little or none of the carbon fiber hull.

I was starting to wonder whether or not the implosion shredded the hull into many small pieces or a few large pieces.

Money shot is around the 10:20 mark. Boom.

Interesting video, but the test isn't really to scale. At depth the pressure doesn't relax as the void is filled. Force due to pressure also increases with the square of surface area. It's not a linear quantity. 6,000 psi on a 2" pipe is not equivalent to 6,000 psi on a 5' void.

That guy got a hell of a YouTube channel though. 6.2-million subscribers now with videos of him crushing stuff with a hydraulic press.

Did you catch the video of the bell jar experiment I posted a few pages back?

That Implosion is closer to what happened.

I missed that post. Funny enough, that vaccum chamber implosion example is from the same creator(s) on youtube as what I posted, just from different channels that they curate (Hydraulic Press Channel and Beyond the press).

I think both approaches are approximations of what actually happened to the sub each with their own limitations. You pointed out the limitations of the pressurized water chamber example pretty well.

The vacuum chamber overcomes the pressure dropping in the pressurized water chamber since the atmosphere provides the reservoir of matter (air) trying to fill the vacuum. The water bath gives a fluid with high density / inertia / momentum characteristics to act on the broken vac chamber. However, that was a glass vacuum chamber (not a carbon fiber tube) and it wasn't a yield failure since he initiated the glass fracture with the shovel. The maximum pressure differential was also limited to 14.7 psi. I don't disagree that the overall vacuum chamber video is a better approximation of a real deep sea implosion.

There was a lot more total energy in play with the real sub than could be represented in either example due to the depth / pressure the Titan was at. Neither video example fully captures that as they don't analyze the hull failure with close to 5000 psi differential (the pressure chamber video shows maximum differential pressures in the 1200 psi range).

I was mostly interested in the performance of the carbon fiber during external overpressure yield failure and the behavior of the carbon fiber / metal head joint during that failure than getting the implosion itself dead-on correct. It's not a perfect approximation, but the video I linked is the only one I've seen with feed from inside the test cylinder using an approximation of the sub (carbon fiber tube with metal caps).

The full investigation will tell the final tale; but, based on the debris videos, my bet is on hoop failure of the carbon tube. That's exactly the failure mechanism shown in the video I linked.

N8Dawg05 said:

---

This is the best similation of the Titan and the forces it was under that I have seen. To summarize, this guy puts a carbon fiber tube with glued on metal end caps inside a water filled chamber and pressurizes it with a hydraulic press until the tube fails.



The carbon fiber tube pancakes upon implosion and blows the end caps off. I think this simulation demonstrates well the forces the sub was under and is consistent with the debris pulled up off the sea bed (titanium caps intact without any meaningful amount if carbon fiber attached). Also may explain where the viewport / window went -> blown out of it's mount as the head was ejected.

I think it is also very telling that one of the titanium rings appears to be elliptical (i.e. it's no longer round) which could indicate that the hull was in the process of imploding as the entire titanium head and ring was blown free.

Whether or not the hull shredded upon implosion is a great question. I would guess that it is in larger pieces but has different buoyance characteristics / was more impacted by undersea currents than the titanium caps and there fore wasn't readily located in the main debris field. Based on what's being said about operating at those depths, you have to get right on top of stuff with lights to identify stuff. It's a darn big ocean floor to search.

---

This scale model test is exactly what I expected. Imploded into a few large chunks and tons of small shards and strands. When carbon fiber fails under its ultimate loading it's quite spectacular.

The one thing that stood out to me was the timescale was a little faster than I expected. The actual sub was a lot larger so had larger distances involved, but based on this scale test, I'd estimate the OceanGate sub fully collapsed within 15 ms.